Introduction to

BioconductorI. Overview of the Bioconductor

Project

Bioinformatics and Biostatistics Lab.,

Seoul National Univ. Seoul, Korea

Eun-Kyung Lee

Outline

What is R?

Overview of the Biocondcutor Project

Microarray

Installing R and Bioconductor

R package structure

Vignette

Sweave

Bioconductor software design

What is R?

A language and environment for statistical

computing and graphics

GNU project

Combined Math and Stat library

Fine graphics

Easy and efficient handling of data

Rich modern statistical routines

Strengths of R

Available on a wide variety of UNIX platforms and

similar systems, windows and MacOS

Easy to produce well-designed publication quality

plots, including math. Symbols and formulae

For computationally-intensive tasks, C, C++, and

Fortran code can be linked and called at run time.

Able to write C code to manipulate R objects

directly

Differences between R and the other statistical SW

R environment

characterize as a fully planned and coherent

system

command line interface (CLI)Preferred for power users

Intimidating for beginners

Longer learning curve

Other SW

an incremental accretion of very specific and

inflexible tools

Graphical user interface (GUI)==> For Novice user, GUI is needed!

Bioconductor Project

An open-source and open-development software

project for the analysis of omics data

R add-on packages

Provide access to powerful statistical and graphical

methods for the analysis

Facilitate the integration of biological metadata from

WWW

Promote high-quality documentation and

reproducible research

Bioconductor Packages

Release 1.9 (Oct. 4, 2006) : 188 packages

Designed for R 2.4.0

Statistical method : cluster analysis, estimation and multiple

testing for linear and non-linear models, resampling,

visualization, etc

Biological assays : cell-based assay, DNA microarray,

proteomics, SAGE, SNP, etc

Biological metadata from WWW : GenBank, GO, KEGG,

PubMed, etc

Interfaces with other languages : C, Java, per, Python, XML,

etc..

Bioconductor Task View : SW

Microarray

one channel, two channel, data input, quality control,

preprocessing, transcription, DNA copy number, SNPs and

genetic variability

affy, marray, multtest, limma, etc.

Annotation

GO, Pathways, Proprietary platforms, Report writing

annotate, AnnBuilder,etc.

Visualization

chromoViz, arrayQCplot, etc

Bioconductor Task View : SW

Statistics

differential expression, clustering, classification, multiple

comparison, time course, sequence matching

limma, qvalue, affylmGUI, timecourse, etc

GraphAndNetworks

GeneTS, Rgraphviz, etc.

Technology

microarray, proteomics, Mass spectrometry, SAGE, Cell

based assays, genetics

Intrastructure

Biobase, Rdbi, tkWidgets, widgetTools, etc.

Microarray

DNA microarray chip

cDNA chip : usually custom based chip

Two channel

One channel

Oligonucleotide chip

Affymetrix

One channel eg) Agilent, Illumina

Central Dogma

cDNA(complementary DNA)

A DNA molecule made in vitro using mRNA as a template and the enzyme reverse transcriptase. A cDNA molecule therefore corresponds to a gene, but lacks the introns present in the DNA of the genome.

cDNA Microarrays

cDNA Microarrays

Affymetrix Microarrays

Affymetrix Microarrays

Microarray

Preprocessing

Background correction

Normalization

Summarization

Affimetrix : affy

Two-channel cDNA : marray

Microarray

exprSet : class for microarray data and methods

for processing them

exprs : the observed expression levels.

annotation : character string identifying the

annotation that may be used for the exprSet

instance

description

notes

phenoData : containing the patient (or case) level

data

Microarray

Analysis

Differential expression

Graph and Networks

Clustering

Classification

Multiple comparison

Time course

Sequence matching

Installing R and Bioconductor

R 2.4.0

Download from CRAN

http://cran.r-project.org

Bioconductor 1.9

Download from Bioconductor website or

http://www.bioconductor.org

From R

> source(http://www.bioconductor.org/getBioC.R)

> getBioC()

Starting and quitting R

Start : R command

Quit : q()

Save : save current env. With save.image

Working directory : getwd, setwd

List objects : ls, objects

Remove objects : rm, remove

Search path : search, attach, detach

Help : help(), ?

R package structure

A structured collection of code, documentation

and/or data

files : DESCRIPTION, INDEX

subdirectories

R

man

doc

src, data, demo, exec, inst

* package.skeleton

How to make R package

Linux/Unix

R CMD check

R CMD build

Windows (need to install a couple of SW)

Rcmd check

Rcmd build

Rcmd build --binary

* http://cran.r-project.org/doc/contrib/cross-build.pdf

How to install and load R package

INSTALL

Linux/Unix

R CMD INSTALL ---.tar.gz

Windows

click

LOAD

library(package.name)

packageDescription()

.find.package()

system.file()

Vignettes

new documentation paradigm

an executable document consisting of a collection of

code chunks and documentation text chunks

provide dynamic, integrated, and reproducible

statistical documents that can be automatically

updated if either data or analyses are changed

Vignettes can be generated using the “sweave”function from the R utils package

Vignettes

Each Bioconductor package should contain at least

one vignette, providing task-oriented descriptions

of the package’s functionality.

located in the doc subdirectory of an installed

package

accessible from the help browser, via the help.start

function

available separately from the Bioconductor website

Vignettes

Biobase package – openVignette function

menu of available vignettes and interface for viewing

vignettes (PDF)

tkWidgets package – vExplorer function

interactive use of vignettes, stepping through code chunks

reposTools package

Sweave

allow the generation of dynamic, integrated and

reproducible statistical documents, intermixing text,

code, and code output(text and graphics)

source file : an executable document consisting of a

collection of code chunks and documentation text

chunks.

utils package : functions

Sweave, Stangle

Sweave

Input (noweb file)

Documentation text chunks

start with @

text in a markup language like Latex

code chunks

start with <<name>>=

R code

file extension

.rnw, .Rnw, .snw, .Snw

Sweave

Sweave function

extract the code chunks, run them and includes their

output(text and graphs) in a .tex file and .ps or .PDF files

Stangle function

concatenates all the code chunks into a .R file

Output (.tex or .pdf)

a single document containing the documentation text, the R

code, the code output (text and graphs)

automatically regenerated whenever the data, code or

documentation text change

Sweave

main.Rnw main.R

main.tex fig.pdffig.eps

main.dvi

main.ps

main.pdf

Sweave

Stangle

latex pdflatex

dvips

Biocouductor SW design

programming approaches used in Biocouductor

Object-oriented S4 class/method framework

to deal with data complexity,

to represent and manipulate various data types

Environments

to provide mappings between different gene

identifies in the annotation metadata packages

closures

for software modularity and extensibility

Experimental Metadata

gene expression measures

scanned image (TIFF)

image quantitation data (.gpr or .CEL )

normalized gene * array matrices of expression measures

( log ratios or summary measures)

reliability/quality information

probe sequence information

information on the target samples hybridized to the

arrays : clinical covariates, experimental condition,

etc.

Experimental Metadata

Standard form

MIAME : minimum information about a microarray

experiment

MAGE-ML : microarray gene expression

Annotation Metadata

Biological attributes that can be applied to the

experimental data

for gene,

chromosome location

gene annotation (LocusLink, GO)

relative literature (PubMed)

Biological metadata sets are large, of different

types, evolving rapidly, and typically distributed via

the WWW

annotate, annaffy, AnnBuilder

Data complexity

large p, small n

dynamic/evolving data

multiple data source : WWW, in-house

multiple data type

quantitative

qualitative

text, graphical

image, sound

censored, missing, erroneous data

various levels of processing

Object-Oriented Programming

adapt OOP paradigm in order to deal with the

complexity of experimental and annotation metadata

S4 class/method design allows efficient and reliable

representation and manipulation of large and

complex biological datasets of multiple types

advantages of class/method design

keep all relevant information in one object

print, summary, accessor/assignment, subsetting, and more

specialized methods

Tools for programming using S4 : methods package

OOP : Classes

provide a software abstraction of a real world

object.

It reflects how we think of certain objects and what

information these objects should contain

defined in terms of slots which contain the relevant

data

An object is an instance of a class

A class defines the structure, inheritance, and

initialization of objects

OOP : Methods and Documentation

Methods

function that performs an action on data

define how a particular function should behave depending on

the class of its arguments

allow computations to be adapted to particular data types

Documentation

special commands can be used to provide and access

documentation for S4 classes and methods, using the type ?

topic syntax.

Methods available for a particular class are listed in the class

help file

OOP : exprSet class

defined in the Biobase package

used to represent processed expression measures

from either Affymetrix or two-color spotted

microarrays

slots

exprs : matrix of expression measures

se.exprs : matrix of SEs for expression measures

phenoData : sample level covariates and responses

description : MIAME information

annotation : name of annotation data

notes : any notes

OOP : phenoData class

defined in the Biobase package

used to keep track of information on target samples

hybridized to the microarray

varLabels : list of variable labels

pData : dataframe of sample level variables

arrays * variables

affy : Affymetrix Oligonucleotide chips

affy package

class definitions for probe-level data and basic

methods for manipulating microarray objects

(printing, plotting, subsetting, class conversions,

etc.)

AffyBatch : probe-level intensity data for a

batch of arrays

ProbeSet : PM, MM intensities for individual

probe-sets

marray: two-channel spotted

microarrays

marray package

class definitions for two-color spotted DNA

microarray data and basic methods for

manipulating microarray objects

marrayLayout : information on microarry

layout

marrayRaw : pre-normalization intensity data

for a batch of arrays (same layout)

marrayNorm : post-normalization intensity data

for a batch of arrays.

pubMedAbst class

annotate package

provide a pubMedAbst class for storing PubMed

abstracts

pmid

authors

abstText

articleTitle

journal

pubDate

Annotation : matching IDs

Accessing annotation information from databases

such as GeneBank, GO, or PubMed, presupposes

the ability to perform the following essential

bookkeeping task

mapping between the different identifiers (IDs)

for a given gene.

one GENENAME

one GenBank accession number

several different GO term IDs

several different PubMed IDs

R Environments

provide key-value mappings

similar to hash tables in other languages

The term key refers to the name of variable,

which can have different values in different

environments.

functions for working with environments include

ls

get, mget(base), multiget(Biobase)

assign(base), multiassign(Biobase)

R Environments

keys can be accessed using

ls(name of the environment)

Values can be accessed using

get(key, envir=name of the environment)

mget(keys, envir=name of the environment)

Closures

consists of the body of the function along with an

enclosing environment containing all variable

bindings needed for evaluating the function.

Closures facilitate software modularity and

extensibility

Summary

Overview of the Biocondcutor Project

Microarray

R package structure

Vignette

Sweave

Bioconductor software design

Next session : focusing on Microarray data analysis